(Lum and Isaac 2016)
(DeVries et al. 2019)
(Gebru et al. 2018)
(Lum and Isaac 2016)
SHAP and
LIME to asses the influence of demographic variables on
model performance.(Lundberg and Lee 2017),(Ribeiro, Singh, and Guestrin 2016)
Consider a binary variable for gender, \(z\) then:
\[ f(\vec{x}) = f(\vec{x}, z) \]
This means: \(P(\hat{y} = 1 | z = 0, y = 1) = P(\hat{y} = 1 | z = 1, y = 1)\)
[Amini et al. (2019)](Shuang et al. 2020)
“Intersectionality means that people can be subject to multiple, overlapping forms of oppression, which interact and intersect with each other.”- Kimberly Crenshaw
(Buolamwini and Gebru 2018)
(Mitchell et al. 2019)
Model explainability is a concept which looks into the ability to understand the results of a machine learning model.
In 2007 a teacher was fired from a Washington DC school due to an algorithm: Despite having highly favourable reviews from students and parents, an opaque algorithm was used to determine her performance as being in the bottom 2% of all teachers.
[Ribeiro, Singh, and Guestrin (2016)][@]
Woohoo! We’ve trained up a model, evaluated that it is working effectively, and completed the R&D process, our models are then deployed into production.